GTC 2020: Building Blocks for Machine Learning Integration into Clinical Workflow
After clicking “Watch Now” you will be prompted to login or join.
Click “Watch Now” to login or join the NVIDIA Developer Program.
Building Blocks for Machine Learning Integration into Clinical Workflow
Krishna Juluru, Memorial Sloan Kettering Cancer Center
Applications of machine learning in radiology image analysis continue to grow at an increasing pace. For these tools to make an impact in diagnostics, they need to be well integrated into a clinical workflow. We'll review the radiology diagnostic interpretation process and the role of several machine-learning algorithms that support radiologists in this effort. We'll then focus on seven generalizable building blocks that are needed to integrate algorithm results into clinical workflow, with roles ranging from quality control and results presentation to error correction and active learning. We'll discuss current standards and the need for new standards, and highlight our experience in applying these algorithms and building blocks in a large cancer center.